The results of a large randomized, controlled trial in France suggest that early antibiotic treatment for bacterial vaginosis does not reduce the risk for late miscarriage or very preterm birth in women who have low-risk pregnancies.
The study of more than 84,000 pregnant women at 40 French hospitals, published in The Lancet, found that the rate of late miscarriage and spontaneous very preterm birth was nearly equal in women with low-risk pregnancies who were randomly assigned to receive clindamycin or a placebo before 15 weeks' gestation. In addition, a sub-trial in women with high-risk pregnancies found no significant difference in rates of late miscarriage or very preterm delivery between women who received either a single course or a triple course of clindamycin.
The authors of the study say their findings indicate that antibiotic treatment for women with no history of late miscarriage or preterm delivery should be reconsidered.
Study finds no risk reduction
Previous research has found that bacterial vaginosis—an excessive growth of certain vaginal bacteria that can result in a major imbalance of vaginal microbiota—doubles the risk of preterm delivery and as much as quadruples the risk when identified before 20 weeks of gestation. Although the mechanisms for how bacterial vaginosis affects pregnancy are unclear, one theory is that it can cause a vaginal infection that travels to the uterus during early pregnancy.
As a result, it has been hypothesized that early antibiotic treatment for bacterial vaginosis might prevent some preterm deliveries. The results of previous clinical trials and meta-analyses, however, have presented a mixed picture on that question.
The purpose of the PREMEVA (Prevention of Very Preterm Delivery by Testing for and Treatment of Bacterial Vaginosis) project, which was funded by the French Ministry of Health, was to test the effectiveness of early clindamycin in reducing the rate of late miscarriages or births before 32 weeks in women with low-risk pregnancies. In addition, the researchers aimed to test the effectiveness of single-course versus triple-course clindamycin in high-risk patients. Clindamycin and metronidazole are the two most frequently used antibiotics for bacterial vaginosis in pregnant women.
From 2006 through 2011, investigators at 40 hospitals screened 84,530 women in the Nord-Pas de Calais region of France for bacterial vaginosis during their first trimester of pregnancy. Of these women, 5,630 were found to have bacterial vaginosis; 3,105 were included in the final analysis.
Overall, 2,869 women with no history of spontaneous preterm birth or late miscarriage were randomly assigned either to one or three courses of oral clindamycin or a matching placebo in a double-blind design. In the high-risk sub-trial, 236 women were assigned to receive either one or three courses of clindamycin. The median gestational age at randomization was 12.4 weeks. The primary outcome was a composite of the onset of late miscarriage (from 16 to 21 weeks) or spontaneous very preterm delivery (between 22 and 32 weeks).
Among the low-risk women, 22 of the 1,904 (1.2%) who received clindamycin had a late miscarriage or very preterm delivery, compared with 10 of the 956 (1.0%) women who received placebo (relative risk [RR], 1.10; 95% confidence interval [CI], 0.53 to 2.32; P = 0.82). Adverse events were reported by 3.0% of women in the clindamycin groups, compared with 1.3% of participants in the placebo group. The most commonly reported adverse event was diarrhea.
In the 236 high-risk pregnancies, the primary outcome occurred in 5 women (4.4%) who received three courses of clindamycin and 8 women (6.6%) who received a single course (RR, 0.67; 95% CI, 0.23 to 2.00; P = 0.47). Adverse fetal and neonatal outcomes did not differ significantly between the two groups.
The authors of the study note that in the only previous randomized trial comparing oral clindamycin with a placebo, researchers found a reduction of two-thirds in the rate of spontaneous preterm deliveries and late miscarriages. Two meta-analyses of studies on antibiotic treatment for bacterial vaginosis in pregnant women, meanwhile, came to differing conclusions: one found that administration of clindamycin before 22 weeks reduced preterm deliveries by 40% and late miscarriages by 80%, while the other found no reduction at all.
The results of this trial, the authors conclude, "provide little evidence that screening and treating all pregnant women with bacterial vaginosis will prevent preterm delivery, and its consequences, when treatment begins before 20 weeks' gestation. Additionally, our findings did not show a reduced risk of late miscarriage."
They add, "In the context of increasing recognition that indiscriminate antibiotic use increases the risk of resistance and might be associated with long-term risks, antibiotic prevention of preterm delivery in women with low-risk pregnancies who have bacterial vaginosis should be reconsidered."
Among the limitations noted by the authors are the compliance rate (80.4% in the clindamycin group completed their treatment vs. 83.7% in the placebo group), the absence of test-of-cure for bacterial vaginosis in the participants, and the low overall prevalence of very preterm delivery in the study population (1.1%).
Results support guidelines
In an accompanying commentary, Mark Klebanoff, MD, of the Center for Perinatal Research at Nationwide Children's Hospital and Rebecca Brotman, PhD, MPH, of the University of Maryland Medical School write that the results of the study are in line with the conclusions of the most recent Cochrane Review on antibiotics for treating bacterial vaginosis in pregnancy.
That review found that antibiotic treatment reduced bacterial overgrowth in more than 7,800 women in 21 clinical trials but did not reduce the number of babies born too early. The Cochrane reviews are considered high quality in evidence-based medicine.
In addition, the results of the new study provide support for guidelines from the US Preventive Services Taskforce and medical-guidance bodies in other countries, which recommend that asymptomatic women who have no history of previous early delivery not be screened or treated for bacterial vaginosis.
"For the majority of pregnant women who have no history of late miscarriage or preterm birth, PREMEVA's results suggest that we have taken our current knowledge of bacterial vaginosis diagnosis, treatment, and preterm birth as far as it can go," they write.
Klebanoff and Brotman conclude that future clinical trials in this population should wait for the creation of advanced molecular tools that can identify which women with bacterial vaginosis are truly at risk for preterm birth, and who might benefit most from antibiotics.
See also:
Oct 12 Lancet abstract
Oct 12 Lancet comment
